A 42-day indoor trial was conducted at the Sustainable Mariculture Laboratory of the Department of Fisheries and Aquaculture, Federal Rural University, Pernambuco Recife, Brazil (08 ^○01’00.16”S, 034 ^○56’57.74”W). The experiment had a 2 × 2 factorial design (monoculture or integrated system with 32% (low) or 40% (high) crude protein content) with the following treatments: IS32 (an integrated system using low protein commercial feed); IS40 (an integrated system using high protein commercial feed); M32 (a monoculture system using low protein commercial feed); and M40 (a monoculture system using high protein commercial feed), all in triplicate. The main ingredients of the feed were soybean meal, wheat bran, meat and bone meal, fish meal, salmon meal, blood meal, viscera meal, rice bran, soy lecithin and fish oil; urea-formaldehyde, propionic acid (preservative) and ethoxyquin (preservative) were also added to the feeds.

Five days prior to stocking, water from an indoor biofloc matrix tank [TAN (total ammonia nitrogen) 0.3 mg/L; NO ₂-N nitrite-nitrogen 0.1 mg/L; NO ₃ nitrate 3.08 mg/L; alkalinity 100 mg CaCO ₃/L, pH 8.1; PO ₄ orthophosphate 2.16 mg/L; TSS total suspended solids 320 mg/L; and salinity 28 g/L] was mixed and equally distributed to fill twelve experimental black-plastic tanks (the useful dimensions of the tanks were 50 × 35 × 23 cm, equivalent to 40 L). The experimental units were maintained under constant aeration by using three cylindrical airstones (diameter 2.4 cm and length 2.6 cm) per tank. No water exchange was carried out during the experimental period, except for the addition of dechlorinated freshwater (2% weekly) to compensate for evaporation losses. Light intensity was maintained at ~ 1000 lux using a fluorescent lamp with a 12 h light / 12 h dark regime.

Molasses were added once a day to maintain the carbohydrate:nitrogen ratio at 12:1. Hydrated lime (Ca(OH) ₂) was added at 10% (by weight) of the daily feed allotment throughout the study to maintain the alkalinity at about 100 mg/L and pH 7.5 (Furtado et al., 2013).

Dissolved oxygen and temperature (YSI model 55, Yellow Springs, OH, USA) were monitored twice a day (at 08:00 and 16:00 h). Salinity (YSI model 30, Yellow Springs, OH, USA) and pH (YSI model 100, Yellow Springs, OH, USA) were monitored twice a week. TAN (Koroleff, 1976), NO ₂-N (Golterman et al., 1978), NO ₃ (Mackereth et al., 1978), TSS (APHA, 2005), PO ₄ ³ (APHA, 2005) and alkalinity (mg/L CaCO ₃) (Felföldy et al., 1987) were monitored once a week.

Specific pathogen-free postlarvae (PL ₁₀, 0.001 g) of L. vannamei were obtained from a commercial laboratory (Potiporã, RN, Brazil). PLs were raised in two 400L rectangular tanks until 20 days (average final weight of 0.23 g), at a shrimp stocking density of 2,500 PLs/m ³ (1,000 shrimp per tank) in salinity of 35 g/L. The cultivation water was prepared by adding Chaetoceros calcitrans, Navicula sp. and Phaeodactylum tricornutum at 15 × 10 ⁴ cells/mL (5 × 10 ⁴ cells/mL for each diatom species) every five days during the culture period. The postlarvae were fed five times a day (at 08:00, 11:00, 13:00, 15:00 and 18:00 h), with a commercial shrimp feed (0.4 to 1 mm in diameter) composed of 40% crude protein and 8% lipids (Aquavita Premiun, Guaraves, Brazil). The daily feeding rate was 35% of body weight at the start of the culture, which was gradually reduced to 20% of body weight after 20 days based on the Van Wyk table (1999).

The experimental units were stocked with Pacific white shrimp L. vannamei (0.23 ± 0.04 g initial weight) at a density of 500 shrimp/m ³ (20 shrimp per experimental unit). The juvenile shrimp were fed three times a day (at 08:00, 12:00 and 16:00 h), with a commercial shrimp feed (Evialis, Presence, Brazil, 1 to 2.5 mm in diameter) ( Table 1). The daily feeding rate was 20% of body weight at the start of the experiment, gradually reduced to 6% of body weight at the end of the 42-day experiment based on the Van Wyk table (1999).

Once a week, 10 shrimp were randomly sampled from each unit and individually weighed on a digital balance (M523 Series Milligram Balances, BEL Engineering, Italy), and put back in the tank. Mean weights were used to determine shrimp growth and adjust the amount of feed and organic carbon offered. At the end of the trial, the total shrimp biomass and the individual shrimp weights from each tank were recorded to determine: final average weight (g) = final biomass (g) / number of individuals at the end; weight gain (g) = final weight (g) – initial weight (g); yield (kg/m ³) = final biomass (kg) / volume of the experimental unit (m ³); survival (%) = (number of individuals at the end of the evaluation period / initial number of individuals) × 100; and FCR = feed supplied / biomass gain.

Samples of Gracilaria biomass were collected at the Pau Amarelo beach, Paulista, Pernambuco, Brazil (07º54’54.74¨S, 034º49’12.07”W), and stored in plastic bags for laboratory analysis. Water was drained from all the samples, the material was carefully inspected to eliminate encrusted organisms and then weighed. Seaweed with reproductive structures, signs of depigmentation and necrosis were discarded (Tsutsui et al., 2010; Marinho-Soriano et al., 2011). Experimental units were stocked with 2.5 kg wet weight of seaweed per cubic meter.

A two-way analysis of variance (ANOVA) was used to determine the effect of the integrated biofloc system and crude protein content (low and high) and their interaction, after confirming homocedasticity (Cochran p ‹ 0.05) and normality (Shapiro-Wilk p ‹ 0.05). Tukey’s test was used when differences between factors and treatments were detected ( p ‹ 0.05). Water quality variables were analyzed by performing repeated ANOVA measures. Data analyses were performed using ASSISTAT Version 7.7 (Assistat Analytical Software, Campina Grande, Paraíba, Brazil).

The water quality variables of temperature (25°C), dissolved oxygen (5.8–6.1 mg/L), salinity (28.7–29 g/L), TAN (0.10 mg/L), NO ₂-N (0.17–0.21 mg/L), NO ₃ (4.21–4.42 mg/L), alkalinity (115.1–136.8 mg/L CaCO ₃), pH (8.1–8.2), PO ₄ (1.99–2.10 mg/L) and TSS (478.4–501.7 mg/L) were not significantly affected ( p › 0.05) by use of a monoculture or integrated system or by the crude protein content ( Table 2).

Shrimp survival rates were all above 80% during the 42-day experimental period and did not differ significantly ( p › 0.05) between the treatments. Growth was more accelerated with the high (40%) crude protein content until 21 days, however, at the end of the experimental period, the final weight was similar in the IS32, IS40 and M40 treatments, with significant effects ( p ‹ 0.05) caused by the use of an integrated system and by differences in crude protein content (Fig. 1 and Table 3).

The yields in the integrated systems were 1.71 and 1.96 kg/m ³, and in the monoculture systems were 1.39 and 1.80 kg/m ³, with significant effects ( p ‹ 0.05) caused by use of the integrated system and differences in crude protein content ( Table 3). The FCR in the integrated systems were 1.47 and 1.67, and in the monoculture system were 1.49 and 1.74, with significant effects ( p ‹ 0.05) caused by the differences in crude protein content (Table 3).

The water temperature (25 °C) was lower than recommended (28 to 32°C) by Van Wyk & Scarpa (1999) for white shrimp culture. This variable influences oxygen consumption, molt cycle, growth, survival, feed consumption rates, metabolic rate and ammonia excretion of shrimp, but did not limit the L. vannamei growth, because the performance variables were considered satisfactory for these conditions. Concerning the water quality variables, mean values of oxygen concentration, salinity, pH, total ammonia nitrogen, nitrite, nitrate and alkalinity were within the range recommended for marine shrimp culture ( Table 2) (Van Wyk & Scarpa, 1999).

Similar results for higher concentrations of nitrate, as compared to nitrite and TAN, were also observed in other studies of integrated biofloc systems (using seaweed and shrimp) (Brito et al., 2016, 2014), thus indicating that nitrifying bacteria were also present in the system. No effect from the use of an integrated biofloc system was observed on dissolved inorganic nitrogen removal, probably due to a lower photoperiod (12:12 light/dark), and lower biomass rate of shrimp:seaweed (1.27 – 1.79) than those used by Sánchez-Romero et al. (2013) (12:12 light/dark and shrimp:seaweed 1:8), and higher TSS levels of about 500 mg/L as compared to Brito et al. (2016), who used 223 to 253 mg/L.

No significant effect was observed on alkalinity and pH levels from the use of an integrated biofloc system or because of crude protein content, because hydrated lime inputs positively influenced alkalinity and pH levels, despite nitrification processes and the conversion of ammonium nitrogen into heterotrophic microbial biomass. No significant effect was observed either on orthophosphate levels from use of an integrated biofloc system, corroborating similar results observed by Brito et al. (2014, 2016). Du et al. (2013) found better phosphate removal by G. asiatica at a 10:1 N:P ratio, however, in this study a lower N:P (2:1) ratio in the water was observed.

Final weight, weight gain, and yield were significantly influenced by use of an integrated biofloc system ( Table 3), while the substitution of a high (40%) for a low (32%) crude protein content commercial feed in integrated biofloc system has no adverse effect on zootechnical variables of shrimp juveniles for these conditions. Correia et al. (2014) and Yun et al. (2016) demonstrated that for L. vannamei reared in biofloc systems, the dietary protein level could be reduced with no adverse effect on shrimp growth and body composition. According to Xu & Pan (2014), a reduction of dietary protein level (35% for 25%) in biofloc systems improves the biofloc proximate composition and the extracellular enzyme activities (amylase), which probably contributes to shrimp growth.

According to Angell et al. (2016), seaweeds have relatively high quality proteins (essential amino acids as a proportion of total amino acids) compared to fishmeal and soybean meal. Moreover, Brito et al. (2016) showed that G. birdiae tissue in a biofloc system had high crude protein levels. These beneficial nutrition components of seaweed may act as nutritional supplements and/or improve the shrimps’ utilization of nutrients and can therefore partially substitute artificial feed (Cruz-Suárez et al., 2010; Gamboa-Delgado et al., 2011; Pallaoro et al., 2016; Peña-Rodríguez et al., 2017).

The data suggest that substituting high-protein (40%) with low-protein (30%) feed in an integrated biofloc system (for 0.2 to 4 g shrimp) can reduce shrimp production costs because of reduced use of crude protein (feed cost). Shrimp growth variables are highly sensitive to protein quality, especially to adequate supply of essential amino acids (Tacon, 1987). Brito et al. (2016) recorded increased crude protein levels in shrimp (in whole bodies) reared in integrated biofloc systems. However, Jatobá et al. (2014) and Yun et al. (2016) did not find significantly different ( p › 0.05) results for protein in whole bodies and muscle tissue in juvenile shrimp reared in a biofloc monoculture system and fed diets containing varying levels of protein.

In addition, Cruz-Suárez et al. (2010) found that shrimps’ fatty acid profile was clearly modified by use of integrated treatments, with a much higher docosahexaenoic acid content in shrimp raised under integrated conditions. This modification is very important because docosahexaenoic acid is part of the cell membrane components of steroid hormone precursors important to homeostatic balance, brain and nervous system tissue, blood clotting and the immune responses (Tacon, 1987).

Other important compounds found in seaweed are β-glucan, carotenoids, tocopherols, polyphenols, polysaccharides - which act as antioxidants and promote bioactivity against virulent antibiotic resistance (Peso-Echarri et al., 2012; Kolanjinathan et al., 2014) - balanced minerals (sodium, potassium, calcium, iron, magnesium, zinc, nickel, copper, cobalt) and vitamins (A, E, C B1, and B2) (Tabarsa et al., 2012; Syad et al., 2013), which may improve shrimp production in integrated biofloc systems. It is important to remember that according to Tacon (1987), reduction in the shrimps’ ability to absorb certain minerals and vitamins from the water may arise under intensive shrimp rearing.

In summary, the results of this study demonstrate that use of an integrated biofloc system can improve feed utilization and growth performance of L. vannamei juveniles. The results also showed that dietary protein content could be reduced to 32% without affecting the zootechnical variables of shrimp juveniles. This suggests that the use of integrated biofloc systems is a sustainable strategy for supporting shrimp growth, as it can allow reducing the crude protein content of their feed. Future studies are needed to fully understand the potential role (protein and lipid retention on shrimp tissue) of seaweed for promoting shrimp nutrition when provided as a supplemental food source in integrated biofloc systems.